Abstract Two Shortcomings in Software Design for Avionics: Timing Analysis and Soft Error Protection
This paper highlights two shortcomings in the current design process of embedded systems of avionics. First, the current software design process does not adequately verify and validate worst-case timing scenarios that have to be guaranteed in order to meet deadlines. Consider the RTCA DO-178B standa...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Text |
| Language: | English |
| Subjects: | |
| Online Access: | http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.87.8421 http://chess.eecs.berkeley.edu/hcssas/papers/Mueller-hcss06.pdf |
| Summary: | This paper highlights two shortcomings in the current design process of embedded systems of avionics. First, the current software design process does not adequately verify and validate worst-case timing scenarios that have to be guaranteed in order to meet deadlines. Consider the RTCA DO-178B standard requiring coverage testing. An additional requirement, namely predictable timing behavior, is essential real-time embedded systems. Airbus requires their suppliers to provide verifiable bounds on worst-case execution time of software for planes under development, Boeing is considering it (e.g., for Airbus 380, Boeing 787 and military aircraft). The automotive industry, among others, is evaluating similar requirements. We provide an analysis of this problem that outlines directions for future research and tool development in this area. Second, the correctness of embedded systems is currently jeopardized by soft errors that may render control systems inoperable. In general, transient faults are increasingly a problem due to (a) smaller fabrication sizes and (b) deployment in harsh environments. In commercial aviation, the next-generation planes (Airbus 380 and Boeing 787) will deploy off-the-shelf embedded processors without hardware protection against soft errors. Since these planes are designed to fly over the North Pole with an order of magnitude higher radiation (due to a thinner atmosphere), system developers have been asked to consider the effect of single-event upsets (SEUs), i.e., infrequent single bit-flips, in their software design. Current developers do not know how to address this problem. We outline much needed research in this area. |
|---|